Poster: Design and Evaluation of 3D Content with ... - Semantic Scholar
Poster: Design and Evaluation of 3D Content with Wind Output Anke Lehmann1, Christian Geiger1, Björn Wöldecke1, Jörg Stöcklein2 1
University of Applied Science Düsseldorf, 2University Paderborn
ABSTRACT In this paper we discuss the effect of two different wind output devices and compare head mounted wind with stationary wind, i.e., wind output of fix mounted fans on a rack. By means of a simple interactive 3D application we evaluated the subjects’ feeling of presence in a pre-test using a standard presence questionnaire. The preliminary results showed that wind output increases presence and indicated a tendency towards stationary wind output. This pilot study delivered us some initial guidance for future in-depth evaluation of wind output. KEYWORDS: 3D Installation, Multimodal Experience, Interaction Device, Haptic Output, Virtual Reality INDEX TERMS: H.5.2 [Information Interfaces and Presentation]: User Interfaces – Haptic I/O, Input Devices and strategies, Prototyping, User-centered design; I.3.6 [Computing Graphics]: Methodology and Techniques – Interaction technique 1
INTRODUCTION
The success of Nintendo's Wii game console or Apple's iPhone is mainly based on its innovative interaction techniques that try to mimic real-world interaction. Less representational interfaces (few or no icons representing objects) and the focus on reality-based interaction styles leverage the user’s built-in abilities by exploiting pre-existing skills and expectations from real-world experiences rather than computer trained skills. Such multimodal interfaces often promise to bring human computer interaction (HCI) closer into the user realm, thereby enabling intuitive, invisible, efficient and robust interaction [7]. Oviatt offers a practical definition of a multimodal system as a system that processes a combination of natural input modes, such as language, touch, haptics, gaze, head and body movements and possibly returns multimedia output [8]. For multimodal 3D/VR applications a potential for intuitive interaction is postulated, even though many realizations seem to demonstrate the opposite due to missing user evaluations. In our projects we work on the usercentred design of multimodal interfaces that support reality-based interaction. In this paper we focus on tactile feedback based on wind output. We extended a simple 3D game and evaluate the influence of different wind output devices. We measured the user’s feeling of presence using a standard questionnaire.
participants indicated a better sense of presence. But in the application user interaction was not possible, the movement occurred with a pre-defined animation path. A head mounted wind display that created wind around the user head was presented 2007 by Cardin, Vexo and Thalmann [2]. The construction used 8 PC fans that were arranged at a distance of 30cm to the user. The fans were connected to the PC via a serial interface and controlled by a VR flight simulator application. The application simulated a wind vector through adjacent fans to be varied by the wind force. Inside of the application the participants should determined the wind direction. The determinations were succeeded relatively with a variation of 8.5 degree. The “Virtual Scooter” project used a vehicle to move within a virtual world [3]. Deligiannidis et al built a wooden scooter as interface that directed a virtual scooter left or right by turning a spring mounted handle bar. Tilting the handle bar was used in a special flight mode. Vibration motors mounted on the handle bar and the running step provided vibro-tactile feedback. A computer-controlled fan simulated head wind, dependent on scooter speed. A button was used to regulate scooter speed. An in-between user test revealed that wind and vibration increases experienced realism. The only wind output display that is available on the market is the AMBX display, a Philips spin-off company [4]. The device is a low-cost game extension featuring two small fans that are placed on the user’s desktop. The fans can be controlled by a small number of 3D games. None of the approaches above compared the presence feeling of wind output using head-mounted devices and fixed mounted devices. 3
3D CONTENT DESIGN AND HARDWARE SETUP
For the evaluation of different wind output devices we implemented a simple 3D skiing computer game. A virtual snowman skates through an ice channel and tries to collect coins while at the same time preventing collisions with trees or a mean penguin. The game was implemented using our HYUI framework that basically extended a commercial 3D authoring system by a number of custom extensions for multimodal feedback and augmented reality tracking [5].
RELATED WORK
2
Moon and Kim [1] showed that the use of wind as output increased the sense of presence. They set up a 1x1x2m room with 20 ventilators in three elevations connected via USB to the computer. The application determined a correct “wind rendering” to simulating wind from any direction with 3-5 ventilators. The 1
2
{anke.lehmann | geiger}@fh-duesseldorf.de, [email protected] Figure 1. Wind output device set-up with 3D application
!EEEEEEEEE SSSyyymmmpppooosssiiiuuummm ooonnn 333DDD UUUssseeerrr !nnnttteeerrrfffaaaccceeesss 222000000999 111444---111555 MMMaaarrrccchhh,,, LLLaaafffaaayyyeeetttttteee,,, LLLooouuuiiisssiiiaaannnaaa,,, UUUSSSAAA 999777888---111---444222444444---333999666555---222///000999///$$$222555...000000 ©©©222000000999 !EEEEEEEEE
!55!
The snowman in the game moves from side to side to avoid collisions. We track the user’s body movement using NaturalPoint TrackIR, a low-cost infrared tracking that we integrated in our system. Visual wind feedback is designed as snow flurry moving from left / right and rendered as particle system that is activated at random and adds a lateral force to the snowman’s movement. Tactile feedback of this force was realized by wind output from fans. To control the different fans we used a DMX dim pack. The digital multiplex (DMX) protocol is most commonly used in this area and we implemented an extension of our framework to control fans with our applications. Two wind output devices were built: a) fixed mounted ventilators and b) head mounted fans. The stand version used two ventilators (150 mm diameter, wind output of 256 m3/h) mounted fix on a traverse. They were positioned 90 cm off the user’s left and right side. The helmeted construction consists of a biker’s helmet with two small PC fans (80 mm diameter, 72 m3/h wind output) each attached on its left and right side. To control the skiing speed we added interaction using the Wii game controller. A simple gesture recognition algorithm for the Wiimote allows us to increase the snowman’s speed if the user performs gestures similar to moving real ski sticks (see figure 1). 4
EVALUATION DESIGN AND PRELIMINARY RESULTS
The sense of presence is a subjective user’s feeling of being involved in a virtual environment. This subjective feeling cannot be measured directly thus an indirect method is used by measure physiological reactions (e.g heart rate, skin response, etc) or by an evaluation with self-reporting scales. We used the latter approach and designed a small experiment in which users tested each experimental condition by playing the game for about 5 minutes. Afterwards the subjects filled out a questionnaire. We used the Igroup Presence Questionnaire (IPQ) by Schubert, Friedman and Regenbrecht [6]. The IPQ gives information about various presence factors, including spatial presence, involvement and realism and calculates an overall presence score. The test environment and the different treatments are illustrated in fig 1. The game was played 3*5 minutes with three different wind feedback types. Condition 1 featured a purely visual representation of snow flurry and was used as baseline. Condition 2 extended the visual representation by wind feedback from ventilators mounted on a traverse. The third condition used two head-mounted fans. We conducted a small pre-test to check the design of our experiment: 12 participants -students and academics from the department of media- participated in this preliminary study (11 m, 3 f., aged from 21 till 31). The task was to move the snowman through the ice canal via body movement. The user tested every version (within-subject experiment) and we permuted the sequence to avoid the influence of learning effects. The Wii controller wasn’t used for the evaluation to provide a simple interaction. Qualitative Analysis: In post interviews the subjects reported that in general they rated the interaction techniques as very intuitive and the wind feedback as increased realism. 75% of the subjects preferred condition 2 and only 25% chose condition 3 as better wind feedback variant. Some subjects complained the missing involvement of the body when using the head-mounted wind device. Also, the vibration and sound of the helmet fans may be another reason of the weaker rating. Quantitative Analysis: The average IPQ presence values for all conditions were analyzed with a one-factor analysis of variance (ANOVA). It shows that there was a statistically significant increase of all presence factors if wind output was used (p
University of Applied Science Düsseldorf, 2University Paderborn
ABSTRACT In this paper we discuss the effect of two different wind output devices and compare head mounted wind with stationary wind, i.e., wind output of fix mounted fans on a rack. By means of a simple interactive 3D application we evaluated the subjects’ feeling of presence in a pre-test using a standard presence questionnaire. The preliminary results showed that wind output increases presence and indicated a tendency towards stationary wind output. This pilot study delivered us some initial guidance for future in-depth evaluation of wind output. KEYWORDS: 3D Installation, Multimodal Experience, Interaction Device, Haptic Output, Virtual Reality INDEX TERMS: H.5.2 [Information Interfaces and Presentation]: User Interfaces – Haptic I/O, Input Devices and strategies, Prototyping, User-centered design; I.3.6 [Computing Graphics]: Methodology and Techniques – Interaction technique 1
INTRODUCTION
The success of Nintendo's Wii game console or Apple's iPhone is mainly based on its innovative interaction techniques that try to mimic real-world interaction. Less representational interfaces (few or no icons representing objects) and the focus on reality-based interaction styles leverage the user’s built-in abilities by exploiting pre-existing skills and expectations from real-world experiences rather than computer trained skills. Such multimodal interfaces often promise to bring human computer interaction (HCI) closer into the user realm, thereby enabling intuitive, invisible, efficient and robust interaction [7]. Oviatt offers a practical definition of a multimodal system as a system that processes a combination of natural input modes, such as language, touch, haptics, gaze, head and body movements and possibly returns multimedia output [8]. For multimodal 3D/VR applications a potential for intuitive interaction is postulated, even though many realizations seem to demonstrate the opposite due to missing user evaluations. In our projects we work on the usercentred design of multimodal interfaces that support reality-based interaction. In this paper we focus on tactile feedback based on wind output. We extended a simple 3D game and evaluate the influence of different wind output devices. We measured the user’s feeling of presence using a standard questionnaire.
participants indicated a better sense of presence. But in the application user interaction was not possible, the movement occurred with a pre-defined animation path. A head mounted wind display that created wind around the user head was presented 2007 by Cardin, Vexo and Thalmann [2]. The construction used 8 PC fans that were arranged at a distance of 30cm to the user. The fans were connected to the PC via a serial interface and controlled by a VR flight simulator application. The application simulated a wind vector through adjacent fans to be varied by the wind force. Inside of the application the participants should determined the wind direction. The determinations were succeeded relatively with a variation of 8.5 degree. The “Virtual Scooter” project used a vehicle to move within a virtual world [3]. Deligiannidis et al built a wooden scooter as interface that directed a virtual scooter left or right by turning a spring mounted handle bar. Tilting the handle bar was used in a special flight mode. Vibration motors mounted on the handle bar and the running step provided vibro-tactile feedback. A computer-controlled fan simulated head wind, dependent on scooter speed. A button was used to regulate scooter speed. An in-between user test revealed that wind and vibration increases experienced realism. The only wind output display that is available on the market is the AMBX display, a Philips spin-off company [4]. The device is a low-cost game extension featuring two small fans that are placed on the user’s desktop. The fans can be controlled by a small number of 3D games. None of the approaches above compared the presence feeling of wind output using head-mounted devices and fixed mounted devices. 3
3D CONTENT DESIGN AND HARDWARE SETUP
For the evaluation of different wind output devices we implemented a simple 3D skiing computer game. A virtual snowman skates through an ice channel and tries to collect coins while at the same time preventing collisions with trees or a mean penguin. The game was implemented using our HYUI framework that basically extended a commercial 3D authoring system by a number of custom extensions for multimodal feedback and augmented reality tracking [5].
RELATED WORK
2
Moon and Kim [1] showed that the use of wind as output increased the sense of presence. They set up a 1x1x2m room with 20 ventilators in three elevations connected via USB to the computer. The application determined a correct “wind rendering” to simulating wind from any direction with 3-5 ventilators. The 1
2
{anke.lehmann | geiger}@fh-duesseldorf.de, [email protected] Figure 1. Wind output device set-up with 3D application
!EEEEEEEEE SSSyyymmmpppooosssiiiuuummm ooonnn 333DDD UUUssseeerrr !nnnttteeerrrfffaaaccceeesss 222000000999 111444---111555 MMMaaarrrccchhh,,, LLLaaafffaaayyyeeetttttteee,,, LLLooouuuiiisssiiiaaannnaaa,,, UUUSSSAAA 999777888---111---444222444444---333999666555---222///000999///$$$222555...000000 ©©©222000000999 !EEEEEEEEE
!55!
The snowman in the game moves from side to side to avoid collisions. We track the user’s body movement using NaturalPoint TrackIR, a low-cost infrared tracking that we integrated in our system. Visual wind feedback is designed as snow flurry moving from left / right and rendered as particle system that is activated at random and adds a lateral force to the snowman’s movement. Tactile feedback of this force was realized by wind output from fans. To control the different fans we used a DMX dim pack. The digital multiplex (DMX) protocol is most commonly used in this area and we implemented an extension of our framework to control fans with our applications. Two wind output devices were built: a) fixed mounted ventilators and b) head mounted fans. The stand version used two ventilators (150 mm diameter, wind output of 256 m3/h) mounted fix on a traverse. They were positioned 90 cm off the user’s left and right side. The helmeted construction consists of a biker’s helmet with two small PC fans (80 mm diameter, 72 m3/h wind output) each attached on its left and right side. To control the skiing speed we added interaction using the Wii game controller. A simple gesture recognition algorithm for the Wiimote allows us to increase the snowman’s speed if the user performs gestures similar to moving real ski sticks (see figure 1). 4
EVALUATION DESIGN AND PRELIMINARY RESULTS
The sense of presence is a subjective user’s feeling of being involved in a virtual environment. This subjective feeling cannot be measured directly thus an indirect method is used by measure physiological reactions (e.g heart rate, skin response, etc) or by an evaluation with self-reporting scales. We used the latter approach and designed a small experiment in which users tested each experimental condition by playing the game for about 5 minutes. Afterwards the subjects filled out a questionnaire. We used the Igroup Presence Questionnaire (IPQ) by Schubert, Friedman and Regenbrecht [6]. The IPQ gives information about various presence factors, including spatial presence, involvement and realism and calculates an overall presence score. The test environment and the different treatments are illustrated in fig 1. The game was played 3*5 minutes with three different wind feedback types. Condition 1 featured a purely visual representation of snow flurry and was used as baseline. Condition 2 extended the visual representation by wind feedback from ventilators mounted on a traverse. The third condition used two head-mounted fans. We conducted a small pre-test to check the design of our experiment: 12 participants -students and academics from the department of media- participated in this preliminary study (11 m, 3 f., aged from 21 till 31). The task was to move the snowman through the ice canal via body movement. The user tested every version (within-subject experiment) and we permuted the sequence to avoid the influence of learning effects. The Wii controller wasn’t used for the evaluation to provide a simple interaction. Qualitative Analysis: In post interviews the subjects reported that in general they rated the interaction techniques as very intuitive and the wind feedback as increased realism. 75% of the subjects preferred condition 2 and only 25% chose condition 3 as better wind feedback variant. Some subjects complained the missing involvement of the body when using the head-mounted wind device. Also, the vibration and sound of the helmet fans may be another reason of the weaker rating. Quantitative Analysis: The average IPQ presence values for all conditions were analyzed with a one-factor analysis of variance (ANOVA). It shows that there was a statistically significant increase of all presence factors if wind output was used (p
